‘Water war’ in the Mekong Basin?

The Mekong River system provides a crucial source of natural resources for riparian nations. However, the increasingly rapid pace of hydro‐development in the Mekong Basin is threatening the integrity of the river system, posing a real concern for Lower Basin states, which are particularly dependent on the basin. This scenario has led to warnings of armed conflict, or even ‘water war’, between riparian states. Certainly, the expanding scale of hydro‐development can be expected to continue increasing interstate tensions in the Mekong region; but are these tensions really likely to escalate to armed conflict? This paper explores this question by drawing on the water and conflict theory of Aaron Wolf. Ultimately, this paper concludes that interstate tensions over Mekong hydro‐development are unlikely to generate armed conflict. This is in part due to the strategic impracticality of such a conflict as well as the presence of a river basin management institution. Most compellingly, though, armed conflict is unlikely because the economic imperative shared by Mekong states is better served by cooperation – or at least non‐interference – than conflict, over regional hydro‐development. In closing, the paper urges that the study of water and conflict in the Mekong Basin be refocused at the intrastate level.     

Identification of α-gliadin genes in <Emphasis Type="Italic">Dasypyrum</Emphasis> in relation to evolution and breeding

To better understand molecular evolution of the large α-gliadin gene family and provide a potential value for wheat quality improvement, total 32 α-gliadin gene sequences were isolated from the two Dasypyrum species, D. villosum. (L.) Candargy and D. breviaristatum (Lindb. F.) Frederisksen. Twelve of 32 sequences contained the in-frame stop condons were predicted to be pseudogenes, suggesting the high variation of gliadin genes in Dasypyrum genome. There are five D. breviaristatum α-gliadin sequences present additional cysteine residues. Four peptides which have been identified as T cell stimulatory epitopes in celiac disease (CD) patients through binding to HLA-DQ2/8 were searched to all Dasypyrum α-gliadin gene sequences, and we found that the distribution of the epitopes varied between Dasypyrum genomes. Phylogenetic analysis of the Dasypyrum α-gliadin genes indicated that the sequences from D. breviaristatum displayed higher variation than those from D. villosum, and the genomic differentiation occurred between the two Dasypyrum species. Moreover, the promoter region of the Dasypyrum α-gliadin genes consisted of four different lengths, indicative of the retrotransposons involving the evolution of the gliadin gene promoters. Based on the specific sequences of the Dasypyrum α-gliadin promoter region, we produced sequence-characterized amplified region (SCAR) markers, and localized the Dasypyrum α-gliadin genes on chromosome 6 VS. The SCAR markers can be used to target the introgression of Dasypyrum α-gliadin genes in wheat–Dasypyrum derivatives.     

The effect of in vivo and in vitro applications of ethrel and GA<Subscript>3</Subscript> on sex expression in bitter melon (<Emphasis Type="Italic">Momordica charantia</Emphasis> L.)

Bitter melon (Momordica Charantia L.) is an important vegetable crop with nutritional and medicinal qualities. As a member of cucurbitaceae it is monoecious with varying proportions of staminate and pistillate flowers. This study was undertaken to determine the effects of various applications of ethrel and gibberellic acid (GA₃) on sex modification in M. charantia. In the first set of experiments, various concentrations of hormones were added to the seed germination medium, in the second, adult plants growing in the field were sprayed with aqueous solutions of ethrel or GA₃ three times at three-day intervals. The number and sex of open flowers was recorded daily for 60 days after the first flower opened and total number of staminate and pistillate flowers was calculated at the end of the experiment. The highest frequency (29.5%) of pistillate flowers was observed in plants treated with 500 ppm ethrel at germination. Similarly, spraying of adult plants with 100 ppm GA₃ increased the proportion of pistillate flowers to 26% relative to 15% in untreated controls. Both ethrel and GA₃ induced significantly higher number of pistillate flowers than control. In vitro hormone application during seed germination was much more successful than spraying of field grown plants.     

The biological characters and polyploidy of progenies in hybridization in 4×–2× crosses in <Emphasis Type="Italic">Dianthus caryophyllus</Emphasis>

Crosses were carried out between tetraploid Dianthus caryophyllus cv ‘Butterfly’ (2n = 4×= 60) and seven diploid cultivars (2n = 2×= 30). Fewer seeds were obtained and the low seed germination was found which suggested the post-fertilization barrier in 4×–2× crosses. 12 progeny were obtained from 5 crossing combinations. Chromosome analysis revealed that they consisted of 5 triploid hybrid plants and 7 tetraploid hybrid plants, suggesting that unreduced male gamete maybe be involved in polyploid formation. Various flower shapes and colours were observed in the polyploid progenies, showing that sexual polyploidization results in greater variability and fitness. The hybrids obtained by 4×–2× crosses showed the flower-size intermediate between the parents or larger than the parents. Some favourable characters of parents such as flower shape, flower colour and resistance to Fusarium oxysporum, were successfully transmitted to the hybrids. Since polyploid hybrids have some of the profitable characters of the parents, they are expected to be used for future breeding in carnation.     

Plant age and in vitro or in vivo propagation considerably affect cold tolerance of Miscanthus × giganteus

The aims of the study were to determine why young Miscanthus × giganteus plants are more frost sensitive during the first winter than older plants, to compare cold tolerance of plants propagated in in vivo and in vitro conditions, and to select plants with higher cold tolerance. The study was performed in three experiments in which plants were prehardened at 12 °C for 2 weeks, hardened at 5 °C for 3 weeks and next chilled at 0 °C or ?3 °C for 3 or 14 days. Afterwards shoot regrowth from rhizomes was evaluated. In Experiment 1 frost tolerance of young plants obtained from a horticultural farm and plants that had already survived the first winter in the field was compared on the basis of LT₅₀ coefficient. In Experiment 2 frost tolerance of plants obtained in vivo and in vitro was compared. Experiment 3 was performed on four groups of plants: in vivo and in vitro obtained plants which were twice selected in cold, as well as in vivo and in vitro obtained plants which were cold treated once. Plants of all these groups were analysed with respect to their frost tolerance. They were prehardened, hardened and subjected to a temperature of 0 °C or ?3 °C for 14 days. The changes in processes accompanying cold acclimation occurring in the rhizomes or leaves of these plants were investigated. The content of abscisic acid, low-molecular antioxidants and phenolics, as well as catalase and non-specific peroxidase activities were analysed.Young commercially obtained plants were more frost sensitive than plants which had survived the first winter in the field. This effect could be caused by a small amount of storage compounds accumulated in finely divided rhizomes produced in a horticultural farm. Prehardening temperature of 12 °C caused more considerable changes in cold acclimation processes in Miscanthus rhizomes than hardening temperature of 5 °C. Plants propagated by in vitro culture were more cold tolerant but only in the first vegetative season compared to plants obtained in vivo. Plants chilled twice demonstrated a higher low-molecular antioxidant level, as well as a greater capability of phenolic accumulation compared to plants which were once cold stressed. Regardless of the recurrence of cold acclimation, ABA level was significantly increased in leaves by prehardening and in rhizomes by hardening. Each repetition of cold acclimation increased cold tolerance and shoot regeneration ability of M. × giganteus rhizomes.     

Genotype × environment interactions in populations possessing <Emphasis Type="Italic">Ga1</Emphasis>-<Emphasis Type="Italic">s</Emphasis> and <Emphasis Type="Italic">ga1</Emphasis> alleles for cross incompatibility in maize

Use of cross incompatibility in corn (Zea mays L.) by the Ga1-s allele may reduce cross-fertilization in specialty and conventional organic corn with pollen from genetically-modified (GM) corn. For effective use, information about environment, and genotype × environment effects on cross-fertilization by ga1 as well as heritability of cross incompatibility in maize is necessary. Our objective was to obtain this information. Four population pairs differing in their genotype at ga1 were evaluated for cross incompatibility with ga1 pollen in different environments. Populations were derived by crossing the recurrent parents B116, PHG35, ARZM16035:S19, and (CHZM05015:Mo17)Mo17 to Ga1-s donor parent Mo508W/Mo506W. Two replicates of each treatment were grown in the center of 952 m² fields planted with purple corn as an adventitious source of ga1/ga1 pollen. Open pollination was allowed and the amount of cross-fertilization estimated by averaging the percentage of purple seeds. Environment and genotype × environment effects were not significant. Contrasts to evaluate differences in cross-fertilization between Ga1-s and ga1 populations revealed that mean percentages of cross-fertilization in Ga1-s populations of B116, ARZM16035:S19, and (CHZM05015:Mo17)Mo17 were significantly lower than in ga1 populations. The estimated broad-sense heritability on an entry-mean basis for cross incompatibility was 0.81. Results suggest differences in genotype at ga1 played a major role in cross-fertilization of populations differing in their genotype at the ga1 locus. Incompatibility may be selected effectively over different environments and the Ga1-s system may be of value to reduce cross-fertilization with GM corn pollen.     

Induction of <Emphasis Type="Italic">Fusarium</Emphasis> wilt (<Emphasis Type="Italic">Fusarium oxysporum</Emphasis> f. sp. <Emphasis Type="Italic">pisi</Emphasis>) resistance in garden pea using induced mutagenesis and in vitro selection techniques

Wilt caused by Fusarium oxysporum f. sp. pisi is a serious production constraint for peas worldwide. An attempt was made to isolate wilt-resistant mutants in two susceptible pea genotypes, Arkel and Azad P-1, employing induced mutagenesis and in vitro selection techniques. Two thousand seeds of each genotype were mutagenized either with ethyl methane sulfonate (EMS, 0.2% and 0.3%) or gamma rays (5-22.5 kR) in ⁶⁰Co gamma cell for three consecutive years. Screening of different mutagenized populations under wilt-sick plots resulted in the isolation of 25 mutants exhibiting complete or enhanced wilt resistance compared to parental genotypes. Five of these wilt-resistant mutants also outperformed the susceptible background genotypes in terms of yield and other horticultural traits. Efforts were also made to isolate wilt-resistant regenerants from callus cultures exhibiting insensitivity to culture filtrate (CF) of F. oxysporum f. sp. pisi. A total of 250 regenerants (R ₀) were obtained from CF-insensitive calli selected from medium supplemented with 20% culture filtrate. When evaluated in artificially inoculated sick plots, only five R ₂ lines obtained from the regenerants exhibited enhanced wilt resistance compared to parental cultivars. However, the selected lines did not exhibit resistance levels equivalent to those shown by wilt-resistant lines isolated through in vivo mutagenesis. To conclude, induced mutagenesis through irradiation and EMS treatments exhibited superiority over in vitro selection for inducing wilt resistance in peas.     

Frequent occurrence of unreduced gametes in <Emphasis Type="Italic">Triticum turgidum</Emphasis>–<Emphasis Type="Italic">Aegilops tauschii</Emphasis> hybrids

Spontaneous chromosome doubling via union of unreduced (2n) gametes has been thought to be the way that common wheat (Triticum aestivum L.) was originated from the hybridization of T. turgidum L. with Ae. tauschii Cosson. Previous works have observed unreduced gametes in F₁ hybrids of Ae. tauschii with six of the eight T. turgidum subspecies. It is not clear, however, whether the formation of these unreduced gametes is a norm in the F₁ hybrids. In the present study, we tried to answer this question by assessing the occurrence frequency of unreduced gametes in 115 T. turgidum–Ae. tauschii hybrid combinations, involving 76 genotypes of seven T. turgdium subspecies and 24 Ae. tauschii accessions. Our data show that these hybrid combinations differed significantly (P ≤ 0.01, F = 11.40) in selfed seedset, an indicator for production of unreduced gametes. This study clearly showed that meiotic restitution genes are widely distributed within T. turgidum. However, significant differences were found between as well as within T. turgidum subspecies and in the interaction of the T. turgidum genotypes with those of Ae. taushii. The possible application of the meiotic restitution genes from T. turgidum in production of double haploids is also discussed.     

Nuclear genome size and chromosome analysis in <Emphasis Type="Italic">Chenopodium quinoa</Emphasis> and <Emphasis Type="Italic">C. berlandieri</Emphasis> subsp. <Emphasis Type="Italic">nuttalliae</Emphasis>

Cytogenetic characterization by karyotyping and determination of DNA content by flow cytometry of seven cultivated varieties of Chenopodium was performed. Chenopodium quinoa cultivar Barandales and C. berlandieri subsp. nuttalliae cultigens Huauzontle, Quelite and Chia roja showed 2n = 4x = 36, x = 9. Statistically insignificant genome size differences for studied varieties ranged from 2.96 pg/2C (1 Cx = 724 Mbp) in C. quinoa to 3.04 pg/2C (1 Cx = 743 Mbp) in Huauzontle. Karyotype analyses revealed the presence of nine groups of four metacentric chromosomes, including two pairs of chromosomes with satellites. The first pair of satellites was located on the largest pair of chromosomes and the second on a different pair of chromosomes in all accessions analyzed. Variation among varieties was evident in chromosome size, genome length (GL) and the position of satellites. Chia roja exhibited greatest GL (58.82 μm) and biggest chromosomes (2.04 μm). Huauzontle showed the smallest GL (45.02 μm) and shortest chromosomes (1.60 μm). Comparison of GL in studied taxa was statistically significant and allowed to define three groups according to the use given to these plants. These data indicate that they are small, very stable genomes in terms of DNA content, and they support the allotetraploid origin(s) of C. berlandieri subsp. nuttalliae and C. quinoa.     

Screening for grain dormancy in segregating generations of dormant × non-dormant crosses in white-grained wheat (<Emphasis Type="Italic">Triticum aestivum</Emphasis> L.)

Pre-harvest sprouting (PHS) in wheat (Triticum aestivum L.) is a significant problem. Introgression of genes controlling grain dormancy into white-grained bread wheat is one means of improving resistance to PHS. In this study seven dormant (containing the SW95-50213 and AUS1408 sources) × non-dormant crosses were produced to investigate the effectiveness of selection for grain dormancy in early segregating generations. Each generation (F₁–F₄) was grown in a temperature controlled glasshouse with an extended photoperiod (i.e. continuous light). F₂ and F₃ generations were subject to selection. Five hundred harvest-ripe grains were tested for germination over a 14 day period, and the 100 most dormant grains were retained and grown-on to produce the next generation within each cross. The response to selection was assessed through analysis of the time to 50% germination (G₅₀) in the F₂, F₃ and F₄ generations. In addition, changes in marker class frequencies for two SSR markers (barc170 and gpw2279) flanking a known quantitative trait locus (QTL) for grain dormancy on chromosome 4A were assessed in DNA from F₂ plants selected from early germinating (non-dormant) and late germinating (dormant) phenotypic extremes within each cross. Selection for grain dormancy in the F₂ and F₃ generations effectively recovered the dormant phenotype in all seven crosses, i.e. the F₄ generation was not significantly different from the dormant parent. Further, selection based on individual F₂ grains changed marker class frequencies for the 4A dormancy QTL; in most cases eliminating the marker class homozygous for the non-dormant alleles. Application of this screening method will enable breeders to better select for grain dormancy and may lead to development of new cultivars offering effective resistance to PHS in the near future.     

Expression of the “<Emphasis Type="Italic">glanded-plant</Emphasis> and <Emphasis Type="Italic">glandless-seed</Emphasis>” trait of Australian diploid cottons in different genetic backgrounds

The expression of the “glanded-plant and glandless-seed” trait was assessed using High Performance Liquid Chromatography (HPLC) analysis methods in different Gossypium hybrids obtained by crossing Australian diploid cottons and various diploid and tetraploid species. Significant variation in the gossypol content in the seed was observed among the analyzed genotypes. HPLC data demonstrated that the gossypol synthesis repression mechanism in the Australian diploid species belonging to C and G genomes was dominant but did not confirm its preferential functioning against A genome species bearing GL₂ locus. About 10% of the produced seeds had total gossypol content lower than the limit imposed by the World Health Organisation (600 ppm) for the use of cotton flour in food and feed. HPLC analysis of the terpenoids aldehyde (TA) contents in the aerial parts of the hybrids showed important qualitative and quantitative variability. This result could indicate a certain separation between pigment gland morphogenesis and terpenoid synthesis mechanisms in cotton.     

Who leads regional industrial dynamics? “New industry creators” in Chinese regions

Recent evolutionary economic geography studies have argued that regional diversification emerges as a path‐dependent process, as regions often branch into industries that are related to its industrial structure. However, it is less clear who are creating new industries and under what regional conditions. This research seeks to fill this gap and identify “new industry creators” in regional industrial diversification. We differentiate two types of new industry formation—path‐breaking and path‐dependent—and examine whether some new industry creators are more path‐breaking than others, by incorporating two factors that have been largely overlooked in recent literature on technological relatedness—firm heterogeneity and regional institutions. Based on a firm‐level data set of China’s manufacturing industries, this paper shows that path‐breaking and path‐dependence coexist. Empirical results confirm that firm heterogeneity and regional institutions not only affect the firms’ capabilities in creating new industries, but also encourage/discourage firms to be adventurous and path‐breaking. This research implies that lagging regions can catch up with developed regions by coordinating regional resources and adjusting local institutional arrangements to attract more path‐breaking firms.     

Winter curing of <Emphasis Type="Italic">Prunus dulcis</Emphasis> cv ‘Butte,’ <Emphasis Type="Italic">P. webbii</Emphasis> and their interspecific hybrid in response to <Emphasis Type="Italic">Xylella fastidiosa</Emphasis> infections

In traditional quantitative genetics, additive effects of genes acting in a population of biparental homozygous lines are estimated on the basis of the phenotypic observations only, usually by taking a difference between mean values for extreme lines. Current molecular methods allow to estimate the additive effects by additionally taking into account the marker data. In this paper we compare these two methods of estimation of additive gene action effects analytically, by simulations and by analysis of real data sets for doubled haploid lines and recombinant inbred lines. The analytic comparison shows under which conditions an agreement of the two methods can be achieved. In most of the considered experimental data and in simulations we observe that the additive effect calculated on the basis of the marker observations is smaller than the total additive effect obtained from phenotypic observations only. This result is discussed, and a weighted regression approach is proposed as a method which can close the gap between the purely phenotypic and genotypic approaches.     

Complex segregation analysis of day-neutrality in domestic strawberry (<Emphasis Type="Italic">Fragaria ×ananassa</Emphasis> Duch.)

Thirty bi-parental progenies were generated using 45 genotypes from the University of California (UC) strawberry breeding population as parents. Both parent genotypes and their offspring were classified for photoperiod insensitivity, or day-neutrality, for flowering using a late-season flowering score and the number of inflorescences per plant recorded during late summer. Complex Segregation Analysis of these traits indicated that their distributions were best explained by a genetic model that postulates a single major locus with partial dominance for day-neutrality in combination with a background of polygenic and environmental variation. The frequency of the allele conferring day-neutrality was estimated as p = 0.59–0.62 in this experimental population. Genotypic values for the inferred major locus were estimated as a = 1.12 and d = −0.81 for the flowering score, and a = 4.93 and d = 2.41 for inflorescence number. Further resolution of inheritance patterns were obtained by comparing the phenotypic variance for each trait with estimates obtained by insertion of these genotypic class values and allele frequencies into standard quantitative genetic models, and by the comparing variance components estimated using a mixed model analysis with and without inferred genotypic classes as a fixed effect. These comparisons suggest that the major gene determines 80.5% and 73.9% of the additive genetic variance for flowering score and inflorescence number respectively. One complicating feature of the results obtained here is that a non-Mendelian model of segregation fit statistically better than a fixed Mendelian model. The genetic parameters estimated using this non-Mendelian model were essentially identical to those obtained with fixed segregation; hybrid and octoploid ancestry, selection affecting flowering response, and limited number of generations in the analysis are discussed as possible explanations of this result.     

科普须防“失真”，“失声”，“失效”

为了提高对科普工作的重视程度,依靠动员科学家推动科普事业,加强科普资源建设,创新科普活动,针对科普工作和活动经常出现“权威性不够”,“针对性不足”,“实效性不强”等问题。根据长期从事农业科普工作的实践,对科普“失真”,“失声”,“失效”现象产生的原因进行了深入思考,提出了解决这些问题的主要对策建议。     

Cytological mechanism of 2<Emphasis Type="Italic">n</Emphasis> pollen formation in Chinese jujube (<Emphasis Type="Italic">Ziziphus jujuba</Emphasis> Mill. ‘Linglingzao’)

In order to elucidate the cytological mechanism of 2n pollen formation in Chinese jujube, a cultivar named ‘Linglingzao’ (2n = 2x = 24) which produces relative more 2n pollens naturally was employed for microsporogenesis analysis. Chromosomes paired in 12 bivalents at diakinesis and the first meiotic division was normal, whereas, the second division was characterized by frequent abnormal spindle orientation (parallel and tripolar spindles). Perpendicular, tripolar, and parallel spindles at metaphase II accounted for 72.93, 17.22, and 9.85% respectively. Perpendicular and tripolar spindles led to tetrads and triads formation respectively. Two types of parallel spindles were observed and a significant association correlation (r = 0.84, P < 0.05) between parallel spindle II (fused spindles) and dyads was found which meant only fused spindles can form dyads, while parallel spindle I with a long distance between two sets of chromosomes seemed to lead tetrad formation. At tetrad stage, the observed frequencies of tetrads, triads and dyads were 80.99, 16.40, and 2.61% respectively. By the rule that each tetrad can form four n pollens, each triad can form one 2n pollen and two n pollens, each dyad can form two 2n pollens, the frequency of 2n pollen estimated was 5.71%. Based on the pollen diameter difference between n and 2n pollens, frequency of 2n pollen observed was 6.15%. No significant difference was observed between frequency of 2n pollen estimated and that of observed. The results indicated that, tripolar spindles and fused spindles at metaphase II followed by more triads and few dyads formation at tetrad stage were responsible for the production of 2n pollen in Chinese jujube. 2n gametes observed correspond to first division restitution. Research on the cytological mechanism of unreduced pollen will provide a platform for unreduced pollen induction and polyploidy breeding in Chinese jujube.     

"种大树"与"大树移植"

由于环境的不断恶化,植树绿化活动在我国城乡风起云涌,形成了热潮。由于“种大树”比栽花种草的生态效益大,所以提倡“种大树”的人越来越多。但目前却有一股“大树移植”风在一些城市劲吹,美其名曰“大树进城”于是十几、二十几厘米胸径的树成千上万地移植进城;更有甚者,将深山老林的古树也挖进城里。笔者认为风不可长。首先应该指出的是“种大树”与“大树移植”是两个不同的概念。虽然“大树移植”也是“种大树”,但提倡“种大树”并不等于提倡“大树移植”。 通常称的“大树”一般指“乔木”而言。《辞海》中“乔木”辞条释为“主干明显直立,分枝繁盛的木本     

Evolution over 16 years in a bulk-hybrid population of subterranean clover (<Emphasis Type="Italic">Trifolium subterraneum</Emphasis> L.) at two contrasting sites in south-western Australia

Evolution was measured over 16 years in a self-regenerating, bulk-hybrid subterranean clover population, consisting of F ₂ seed from 253 crosses, sown at Nabawa and Mt Barker, short and long growing season sites, respectively, in south-western Australia. Seed bank samples harvested annually were grown in a common garden. Experiment 1 measured flowering time in plants from each year, while Experiment 2 measured 26 variables in the populations 3 and 16 years after sowing, in comparison with the ancestral population. Changes in population means were observed in 20 characters and variability declined in 11 characters at one or both sites, with much of this occurring within the first three years. Natural selection at Nabawa favoured early flowering of long duration, thick peduncles, high harvest index and high hardseededness. At Mt Barker it favoured late flowering of short duration, large leaves and long, thick petioles at flowering, thick stems with long internodes, long, thin peduncles with a high burial angle, large plants at maturity, low hardseededness and high biochanin A and total isoflavone contents. High seed production capacity, with high seed weight and seeds per burr, was important at both sites. The use of bulk-hybrid populations is demonstrated as a low-input means of breeding and selecting well-adapted subterranean clovers.

Re‐Regulation in the Post‐WTO Period? A Case Study of Vietnam's Food Retailing Sector

Academic interest on domestic regulatory (and re‐regulatory) impacts of retail foreign direct investment remains surprisingly under‐researched, despite high‐profile campaigns, particularly in Southeast Asia, to rein in the expansion of retail transnational corporations. This paper focuses on the trends of re‐regulation of foreign retailers, particularly in the food sector, in Vietnam before and after the accession to the World Trade Organization (WTO) in 2007. The findings reveal a complex layering of regulation, some of it a holdover from the pre‐2007 period and some of it occurring since WTO accession, in the form of the controversial Economic Needs Test.     

Analysis of endosperm development of allotriploid × diploid/tetraploid crosses in <Emphasis Type="Italic">Lilium</Emphasis>

Lilium are Fritillaria-type plants. Triploid lilies, regardless of their male sterility, can be used as female parents to cross with appropriate males, in contrast to other triploid Polygonum-type plants, which are usually seedless. Up to now, however, little attention has been paid to the reason. Here we made allotriploid × diploid/tetraploid crosses in Lilium. Endosperm and progenies of LAA × AA/AAAA crosses were analyzed for ploidy level and genome composition. The endosperm of some LAA × AA/AAAA ovules developed well but there was little or no development of LAA × LL/OO endosperm. The endosperm genome composition of LAA × AA, LAA × AAAA, LAA × LL, LAA × OO are theoretically derived as 5A + 2L, 6A + 2L, 4A + 3L, 4A + 2L + O, respectively. Genomic in situ hybridization showed that the progenies of LAA × AA/AAAA were aneuploid. Based on the experimental results and analyses, our hypothesis is that five same genomes of endosperms are essential for its development in triploid × diploid/tetraploid crosses of Lilium. The hypothesis can explain the success or failure of 3x × 2x/4x crosses in Lilium and is of importance for lily breeders who need to know the likelihood of success when producing new cultivars of this economically important horticultural crop.